Theoretical foundation of a decision network for urban development

Planning problems are challenging and complex in that they usually involve multiple stakeholders with multi-attribute preferences. Thus few, if any, planning tools are useful in helping planners address such problems. Decision analysis is less useful than expected in dealing with planning problems because it focuses overwhelmingly on making a single decision for a particular decision-maker. In this paper, we describe the theoretical foundation of a planning tool called the ＇decision network＇, which aims to help planners make multiple and linked decisions when facing multiple stakeholders with multi-attribute preferences. The research provides a starting point for a fully fledged technology that is useful for dealing with complex planning problems. We first provide a general formulation of the planning problem that the decision network intends to address. We then introduce an efficient solution algorithm for this problem, with a numerical example to demonstrate how the algorithm works. The proposed solution algorithm is efficient, allowing computerization of the planning tool. We also demonstrate that the diagrammatic representation of the decision network is more efficient than that of a decision tree. Therefore, when dealing with challenging and complex planning problems, using the decision network to make multiple and linked decisions may yield more benefits than making such decisions independently.

规划问题具有挑战性和复杂性,因为它们通常涉及多重属性偏好和多个利益相关者。因此,很少有规划工具能帮助规划者解决这些问题。传统决策分析在处理规划问题时无法达到预期,因为它主要侧重于为特定决策者制定单一决策。本文介绍了一个名为"决策网络"的规划工具的理论基础,其目的是帮助规划者在面对具有多属性偏好的多个利益相关者时,作出多重且相互关联的决定。本研究提供了一个成熟科学技术的起点,将有助于处理复杂的规划问题。我们首先提出决策网络模式所欲处理的规划问题的一般化模式。然后我们针对这个规划问题提出一个有效的解决演算法,再使用数值示例来演示该演算法是如何运作。所提出的解决演算法被证明是有效的,亦使该规划工具的计算机化成为可能。同时,我们指明"决策网络"的图形表示较"决策树"的图形表示来得更为有效。因此,在处理具有挑战性和复杂性的城市规划问题时,通过决策网络模式所作出的多重的且相关的决策,将会比独立地做出这些决策产生更多的好处。